This invention relates to improvements in fault monitoring devices for computer controlled systems. More particularly, it relates to improvements in watchdog timers, or fault watch timers, which permit their use with a series connected light emitting diode, LED, indicator. The basic timing function is described in Disclosure Document No. 160368, filed Dec. 11, 1986, in the U.S. Patent and Trademark Office.
Computer control of processes is achieved by programming various output signals which are transmitted to the controlled device from the central processing unit, CPU. In most instances, these control signals toggle between on and off states or cause reversal of state in the controlled device. These state changes are periodic and can be monitored by a device such as a watchdog timer, or fault watch timer, which are currently commercially available and are commonly used in computer controlled operations.
The watchdog function is usually accomplished by an edge-triggered change-of-state detector which resets a timer having a predetermined time-out period So long as the programmed state changes continue, the timer is continually reset, and the watchdog timer output is inactive. However, if computer control is lost due to program error, brown-out, blackout, electromagnetic interference, radio frequency interference, or other cause, the expected change-of-state will not occur, and the watchdog timer will time-out and activate its output to alarm, reset, or shut down the system. All currently available watchdog timers function in essentially this same manner.
Typical prior art watchdog timer circuits are intended to obtain power from a direct power supply connection. Such circuits are not suitable for operation in an Input/Output mounting rack where all circuit power is by means of an LED in series with the power supply, because of excessive or insufficient current draw of these types of watchdog timers, in addition to their minimum voltage requirements.